Obtain And Process The Point Cloud Based On Image Processing

Reverse engineering with unique directly from physical model to build a computer modelmethod for CAD technology revolution, involving3D modeling and data visualization,machine vision and the key technology such as3D cognitive, in education, scientific research,entertainment, real estate, tourism, exhibition, advertising, engineering design andmanufacturing and other industries have a wide range of applications.From the physical model of data generally exists in the form of point cloud, point clouddata is very large, usually reached levels millions or even billions. Quality and amount ofpoint cloud data is point cloud computer3D model reconstruction of the two key problems ofquality.Based on the analysis of the existing on the basis of the point cloud to acquire advancedtechnology, the active3D point cloud based on binocular vision method, study on the imageprocessing and the point cloud processing. Experimental apparatus includes a projector, twoindustrial cameras and a personal computer.Paper analyzes industrial camera, projector and the relationship between the quality ofimage acquisition, image acquisition, image noise based on wavelet transform to removebackground information and based on the extracted contour image problems, the experimentalresults show that the image processing work effectively reduces the image quality brings tothe point cloud computing error.On the basis of summarizing predecessors’ work, the paper studies a kind of contourmatching and mutual combination of pixel level matching and adaptive matching algorithm.The algorithm based on gray code images, each gray code value is the same as a condition ofcomparison, the number of matching point position is calculated. The experimental resultsshow that the algorithm improves the matching precision effectively.In point cloud processing process, this paper uses the mathematical morphology methodto deal with the noise in point cloud; the grid method is adopted to simplify the processing, point cloud data clustering method is used to segmentation of point cloud data. Among them,the point cloud simplification method can on the basis of the basic details to keep the originaldata, greatly reduce the amount of point cloud data.